Stefan Wiktor, MD, MPH
Chief, Surveillance and Infrastructure Development Branch,
Global AIDS Program, CDC
Background: Access to HIV testing is a critical component of HIV prevention and care services. Because of their ease of use, HIV rapid tests are being widely used to provide HIV testing services in resource-poor countries, and in-country evaluations and quality control of point-of-service testing are important steps to assure the quality of this testing. To date there is limited information concerning the outcomes of these evaluations and of the quality of rapid test performance when used in clinical settings.
Methods: A standardized questionnaire regarding the evaluation and implementation of HIV rapid tests was administered to personnel in ministries of health and reference laboratories in 10 African countries where the U.S. Centers for Disease Control and Prevention has provided technical assistance to improve laboratory capacity. Additional information was obtained from reports, presentations, personal communications, and other documentation. The existing HIV testing algorithms in each country served as the gold standard for measuring the performance of each rapid test.
Results: Results are available from HIV rapid test evaluations conducted between 2001 and 2003 in reference laboratories in 11 countries. In all, 18 different rapid tests were evaluated, and the median sensitivity was above 99% for 12 of these tests (range 92.5% – 100%), and the median specificity was above 99% for 14 of the 18 rapid tests (range 95.3% – 100%). Subsequent evaluations at point-of-service sites in 5 countries demonstrated sensitivities and specificities of rapid tests comparable with the results of reference laboratory evaluations. To date, rapid testing has been implemented in 8 of the countries in over 600 sites including voluntary counseling and testing centers, antenatal clinics, blood donation centers, hospitals, and clinics. A median of 96,000 people were tested in each country in 2003 (range: 15,000 to 315,000 people). A serial (consecutive) testing algorithm was used in 5 countries and a parallel testing algorithm was used in 3 countries. Quality assurance systems in four countries involve retesting a sample of specimens at a reference laboratory. Concordance between on-site rapid testing and retesting at the reference laboratory ranged from 95.7% to 99.5% (median: 98.7%).
Conclusions:In-country evaluations of HIV rapid tests in settings with varying levels of technical skill and different HIV-1 subtypes yield results comparable to those obtained in more controlled evaluations. Furthermore, the excellent validity of rapid testing algorithms at point-of-service sites argues for wider implementation of HIV rapid testing.

Last Update: April 8, 2005